Join Zoom Meeting

https://hku.zoom.us/j/97837967715?pwd=dam4MZt8SyqQBYvhKtuH7b0imUvzZj.1

Meeting ID: 978 3796 7715

Password: 919879

Abstract:

Set-based approaches are highly relevant for controller synthesis, state estimation and formal verification of dynamic systems, since many tasks require handling worst-case uncertainties and establishing reliable boundaries for system behaviors. Compared with approaches relying on stochastic distribution of uncertainties, set-based approaches are more robust and accessible. As the common foundation of these tasks, set representations directly determine the accuracy, the efficiency and the space overhead of set-based approaches. In order to choose a proper set representation for practical tasks, one needs to comprehensively consider the set representability, computational complexity, and whether such sets are closed under targeted set operations such as linear mapping, Minkowski sum, and intersection. In this seminar, we will review the advantages, the challenges, and the applications of traditional and emerging convex set representations including polytopes, zonotopes, ellipsoids, constrained zonotopes and ellipsotopes, as well as introduce the latest advances of convex set representations.

ALL INTERESTED ARE WELCOME